[elephant-devel] Re: Derived Indicies

[Alex Mizrahi]

 ??>> The derived index hack is still more efficient for large sets.
 ??>> Without changes to the data stores to create an efficient way of
 ??>> sorting concatenated values, I don't see a way to improve on it
 ??>> easily.

 LPP> I'm not sure you actually need concatenated index values at all
 LPP> if you manage your objects correctly. I.e. putting them in appropriate
 LPP> containers (the natural OODB way) as opposed to throwing them all
 LPP> together in some indexing namespace and then tediously (for programmer
 LPP> and machine) selecting the stuff you need.

actually tuple keys are very similar to multiple namespaces.
with tuple keys we have essentially a table of triples (userid, time, 
message).
for specific value of userid, you have a well-defined section of this table, 
that essentially represents inbox of that user.

you can as well call userid "a separate namespace" and call that separate 
section "a separate btree", but that won't change much.